COMPLEX STRESS DISTRIBUTION IN ENGINEERING MATERIALS. 341 
trations will have pronounced effects. There is, however, little available information 
to enable one to decide if a material with a drop of stress at yield possesses any advan- 
tage over other materials which have not this property. According to Stanton,® who 
tested various samples of steel having sharp corners and therefore stress concentrations, 
a steel with a low carbon content realises a larger percentage of its maximum resistance 
than does steel with a higher carbon content. The lower carbon steel is the one which 
will have the greater drop of stress at yield. It thus appears that in this connection 
also the drop of stress at yield is indicative of a property which it is highly desirable 
that materials of construction should possess. It is hoped that further experimental 
work on the fatigue of specimens with stress concentrations will be undertaken. 
REFERENCES. 
1 See R. and C. Proceedings Roy. Soc., p. 463. 
2 See R. and C. Proceedings Roy. Soc., p. 465. 
3 Southwell, Engineering, August 23, 1912. 
4 University of Illinois. Bulletin 68. 
5 B.A. Stress Committee Report. 1923. 
6 Stanton, Engineering, April 19, 1907. 
TABLE V. 
Uniform Bending 
, Approx. ; fr 
Mark Section ; : Compression 
Pineeusiine iii Uietind! Wuodateaee fe 
Limit Rupture, fr fe 
Ibs. O” | lbs. 0” | lbs. 0” 
1 | 3” x 1:5" 4,550 | 6,560 4,200 1-56 
2 i | of 4,550 6,560 4,640 1-42 
3 aGf 0-8” | 3 5,000 5,550 4,500 1-23 
4 a 0-4” | cf 5,300 5,600 5,110 1-10 
1 R. V. Southwell, Hngineering. 
2 Strength of I Beams by Flexure, Bulletin 68, University of Illinois. 
3 B.A. Stress Committee, 1923. 
4 Stanton, Hngineering, April 19, 1907. 
VIII. Note on Impact Experiments. 
By R. V. Sournwett, M.A. 
In the progress of engineering science, a steady improvement in the quality of 
the materials which the metallurgist can supply is accompanied by an equally con- 
tinuous increase in the stringency of the demands which are made upon them by the 
designer ; and as a consequence new tests come to be needed, for ensuring that a 
material does in fact possess the properties desired. But the problem of devising a 
suitable test is, in general, far from simple—not only because it is difficult to form an 
exact notion of the property in question (as is shown by the vagueness which still 
attends the definition of ‘ hardness,’ ‘ toughness,’ ‘ ductility,’ and the like), but also 
because it is almost impossible to achieve conditions of testing such that this property 
has a predominant influence on the results. Hence it comes about that many tests 
are current in engineering practice to-day which, although they serve in a general 
way to discriminate between ‘sound’ and ‘unsound’ material of a given class, are 
very difficult to interpret with precision, as indicative of some definite physical quality. 
